# Dolphin > Uncensored open models and a decentralized network that serves them on idle consumer GPUs. Canonical: https://www.subconscious.dev/providers/dolphin · By The Subconscious Team · Updated October 11, 2026 - Example models: DeepSeek V4.1 Flash, Dolphin Mistral 24B - Website: https://dphn.ai ## Overview Dolphin began in 2023 as Eric Hartford's series of open-source, uncensored fine-tunes, trained to drop the refusals and baked-in guidelines of their base models so that the developer sets the rules through the system prompt. Current releases include Dolphin Llama 3.1 8B and Dolphin Mistral 24B Venice Edition, built with Venice, and the project publishes nearly 100 models on Hugging Face. It also runs a chat app, a Telegram bot and a synthetic data service that generates answers to commonly refused prompts. Dolphin Network is its inference layer: gamers and other GPU owners run node software while their cards sit idle and earn $POD tokens for the work. Validators sample requests and check returned logprobs against a trusted copy of the model, so nodes cannot quietly swap in a smaller model, and nodes cannot keep logs. The pitch is price, since most operators already own their hardware. The OpenAI-compatible Dolphin API launched in October 2026 with DeepSeek V4.1 Flash at $0.01 in and $0.07 out per million tokens. Signing up means connecting a crypto wallet and authorizing through Telegram. ## Upsides - Very low prices on DeepSeek V4.1 Flash. - Uncensored models and a network built to host models mainstream clouds will not. - Requests are encrypted end to end and nodes cannot log them. ## Core use cases - Cheap batch or background calls where tail latency matters little. - Products that need models without vendor-imposed refusals. ## Downsides - API is brand new, with one hosted model and no published speed, context or SLA. - Runs on consumer GPUs across the internet, so latency and uptime are less predictable. - Wallet-and-Telegram signup and a crypto token will not clear most procurement teams. ## At a glance | Attribute | Value | |---|---| | Model access | Open weights, uncensored fine-tunes | | Flagship models | DeepSeek V4.1 Flash, Dolphin Mistral 24B | | Speed | - | | Price | $0.01 in, $0.07 out (DeepSeek V4.1 Flash) | | Customization | Own uncensored fine-tunes | | Deployment | Decentralized serverless API | | Long context | - | ## FAQ ### What is Dolphin? Dolphin began in 2023 as Eric Hartford's series of open-source, uncensored fine-tunes, trained to drop the refusals and baked-in guidelines of their base models so that the developer sets the rules through the system prompt. Current releases include Dolphin Llama 3.1 8B and Dolphin Mistral 24B Venice Edition, built with Venice, and the project publishes nearly 100 models on Hugging Face. It also runs a chat app, a Telegram bot and a synthetic data service that generates answers to commonly refused prompts. ### What is Dolphin best for? Cheap batch or background calls where tail latency matters little; Products that need models without vendor-imposed refusals. ### How much does Dolphin cost? Dolphin pricing at a glance: $0.01 in, $0.07 out (DeepSeek V4.1 Flash). Rates change often, so check Dolphin's pricing page before committing. ### What are the downsides of Dolphin? API is brand new, with one hosted model and no published speed, context or SLA; Runs on consumer GPUs across the internet, so latency and uptime are less predictable; Wallet-and-Telegram signup and a crypto token will not clear most procurement teams. ### What are the best alternatives to Dolphin? Common alternatives include Subconscious, OpenAI, Anthropic, Google Vertex AI, Amazon Bedrock. Each has a head-to-head comparison with Dolphin on this site. ## Comparisons - [Subconscious vs Dolphin](https://www.subconscious.dev/compare/subconscious-vs-dolphin.md) - [OpenAI vs Dolphin](https://www.subconscious.dev/compare/openai-vs-dolphin.md) - [Anthropic vs Dolphin](https://www.subconscious.dev/compare/anthropic-vs-dolphin.md) - [Google Vertex AI vs Dolphin](https://www.subconscious.dev/compare/google-vertex-vs-dolphin.md) - [Amazon Bedrock vs Dolphin](https://www.subconscious.dev/compare/aws-bedrock-vs-dolphin.md) - [Together AI vs Dolphin](https://www.subconscious.dev/compare/together-ai-vs-dolphin.md) - [Fireworks AI vs Dolphin](https://www.subconscious.dev/compare/fireworks-vs-dolphin.md) - [Baseten vs Dolphin](https://www.subconscious.dev/compare/baseten-vs-dolphin.md) - [Groq vs Dolphin](https://www.subconscious.dev/compare/groq-vs-dolphin.md) - [Cerebras vs Dolphin](https://www.subconscious.dev/compare/cerebras-vs-dolphin.md) - [DeepInfra vs Dolphin](https://www.subconscious.dev/compare/deepinfra-vs-dolphin.md) - [Hugging Face Inference Providers vs Dolphin](https://www.subconscious.dev/compare/hugging-face-vs-dolphin.md) - [Modal vs Dolphin](https://www.subconscious.dev/compare/modal-vs-dolphin.md) - [Cloudflare Workers AI vs Dolphin](https://www.subconscious.dev/compare/cloudflare-workers-ai-vs-dolphin.md) - [xAI vs Dolphin](https://www.subconscious.dev/compare/xai-vs-dolphin.md) - [Mistral AI vs Dolphin](https://www.subconscious.dev/compare/mistral-ai-vs-dolphin.md) - [DeepSeek vs Dolphin](https://www.subconscious.dev/compare/deepseek-vs-dolphin.md) - [Moonshot AI vs Dolphin](https://www.subconscious.dev/compare/moonshot-ai-vs-dolphin.md) - [Z.ai vs Dolphin](https://www.subconscious.dev/compare/z-ai-vs-dolphin.md) - [Alibaba Cloud vs Dolphin](https://www.subconscious.dev/compare/alibaba-cloud-vs-dolphin.md) - [Meta vs Dolphin](https://www.subconscious.dev/compare/meta-vs-dolphin.md) - [Cohere vs Dolphin](https://www.subconscious.dev/compare/cohere-vs-dolphin.md) - [SambaNova vs Dolphin](https://www.subconscious.dev/compare/sambanova-vs-dolphin.md) - [Nebius vs Dolphin](https://www.subconscious.dev/compare/nebius-vs-dolphin.md) - [Crusoe vs Dolphin](https://www.subconscious.dev/compare/crusoe-vs-dolphin.md) - [fal vs Dolphin](https://www.subconscious.dev/compare/fal-vs-dolphin.md) - [Novita AI vs Dolphin](https://www.subconscious.dev/compare/novita-ai-vs-dolphin.md) - [Venice vs Dolphin](https://www.subconscious.dev/compare/venice-vs-dolphin.md) - [Parasail vs Dolphin](https://www.subconscious.dev/compare/parasail-vs-dolphin.md) - [Inference.net vs Dolphin](https://www.subconscious.dev/compare/inference-net-vs-dolphin.md) - [GMI Cloud vs Dolphin](https://www.subconscious.dev/compare/gmi-cloud-vs-dolphin.md) - [Thinking Machines vs Dolphin](https://www.subconscious.dev/compare/thinking-machines-vs-dolphin.md) - [Sail Research vs Dolphin](https://www.subconscious.dev/compare/sail-research-vs-dolphin.md) - [Morph vs Dolphin](https://www.subconscious.dev/compare/morph-vs-dolphin.md) - [Relace vs Dolphin](https://www.subconscious.dev/compare/relace-vs-dolphin.md) - [TypeSafe AI vs Dolphin](https://www.subconscious.dev/compare/typesafe-ai-vs-dolphin.md) - [StepFun vs Dolphin](https://www.subconscious.dev/compare/stepfun-vs-dolphin.md) - [Runware vs Dolphin](https://www.subconscious.dev/compare/runware-vs-dolphin.md) - [StreamLake vs Dolphin](https://www.subconscious.dev/compare/streamlake-vs-dolphin.md) - [Wafer vs Dolphin](https://www.subconscious.dev/compare/wafer-vs-dolphin.md) - [RunInfra vs Dolphin](https://www.subconscious.dev/compare/runinfra-vs-dolphin.md) - [Particle.AI vs Dolphin](https://www.subconscious.dev/compare/particle-ai-vs-dolphin.md) - [Luminal vs Dolphin](https://www.subconscious.dev/compare/luminal-vs-dolphin.md) - [Infron vs Dolphin](https://www.subconscious.dev/compare/infron-vs-dolphin.md) - [Inception vs Dolphin](https://www.subconscious.dev/compare/inception-vs-dolphin.md) ## Sources - [Dolphin](https://dphn.ai/) - [Dolphin Network](https://dphn.ai/network) - [Dolphin API quick-start and pricing](https://dphn.ai/docs/coupon-credits/) - [Dolphin models](https://dphn.ai/models) Pricing and model lineups change often; figures are a snapshot.